Differences in procedural (skill-based) learning may be highly relevant to autism, with several investigators positing that impaired development of social and communicative skills may be linked to anomalous formation of action models beginning early in infancy. Consistent with this idea, there is increasing evidence that autism is associated with anomalous motor development, including impaired imitation and execution of goal-directed actions (dyspraxia). During the initial funding period we therefore began a research program focused on studying procedural learning at a level of detail best achieved through careful examination of motor learning. Our studies led to a critical observation: there is a fundamental difference in how children with autism learn motor skills: they build a stronger than normal association between motor commands and intrinsic proprioceptive feedback and a weaker than normal association between the same commands and extrinsic visual feedback. The findings not only help explain impaired skill development in autism, they may also explain why individuals with autism develop a bias toward their own inner world and struggle to understand and interpret the meaning of others' social and communicative actions. This excessive reliance on proprioceptive feedback to guide motor learning may have its roots in the wiring of the brain: imaging and post-mortem studies have revealed an over-expression of short-range axons, including those connecting neighboring primary motor cortex and somatosensory cortex, where proprioception is encoded. Our imaging findings from the initial funding period provide preliminary support for this hypothesis, revealing motor skill impairment in autism is associated with increased volume and paradoxically decreased organization of short-range white matter connectivity in primary sensorimotor cortex and decreased functional connectivity between distant brain regions. Clearly, questions remain regarding the clinical relevance and neural basis of the stronger than normal association between motor commands and proprioceptive feedback in autism and whether a change in this bias can be achieved, providing an effective method of treating impaired skill acquisition. Consequently, we propose to: confirm that children with autism show a bias towards proprioceptive-guided motor learning and determine whether this is associated with the core social, communicative, and motor skill deficits that define the disorder (Aim 1); use both structural (Aim 2) and functional (Aim 3) MRI methods to examine the neural basis of autism-associated alterations in motor learning; and working toward a treatment, determine whether transcranial direct current stimulation (tDCS) can help children with autism develop internal models of behavior that are better at predicting external visual consequences, as is important for guiding socialization and communication (Aim 4). The proposed studies will establish a foundation for using advanced methods of motor and imaging analysis to examine the brain basis of autism, and, most critically, for designing and testing novel methods of treatment. PUBLIC HEALTH RELEVANCE: Children with autism have particular difficulty with development of social and communicative skills, as well as motor skills. Skill development depends on procedural learning mechanisms that are best examined through careful study of motor skill learning, a process that also appears to be important for learning how to understand the meaning of other people's actions. The purpose of this research is to identify the brain mechanisms in autism that contribute to impaired acquisition of motor skills, and by extension social skills, and to identify methods for improving how children with autism learn these skills.